#WritersCoffeeClub#Writing May 13: When does AI cross the line between helpful tool and problematic tech for writing and publishing?
I am an active luddite and do what I can to sabotage AI.
When does it cross the line?
When steals training material.
When it plagiarizes or mimics others.
When they utilize immoral terms of service. i.e looting social media that have implemented "legalized theft."
When it attempts to substitute itself for real individuals.
When it attempts to pass itself off as human output.
When it attempts to imply it is "intelligence."
In other words, everything that currently labels itself as AI. "Deep Learning" tools could be acceptable but not as they are currently implemented and run.
@NaraMoore Actually, it's never the ML model that does all these things. We're not quite at that level of sophistication (or lack of ethics) yet. It's always humans who try to sell you something. Ok, sometimes it may produce lies, but nobody should expect you to believe them.
The dominant AI models are trained on stolen material or utilize the "legalized theft" implemented by SNS platform's TOS. Sweatshop labor is used to groom it. It does not credit its sources which is plagiarism.
It does nothing to hinder unethical people from passing off the products as their own. So while it may not directly pass itself off as human-produced it blandly winks at it.
@NaraMoore I meant, we're not quite at the point yet where we let ML replace CEOs, marketing, legal, development, security, etc.
It's still some humans making the (bad) decisions.
I am all for banning guns AND suing gun manufacturers.
I am all for poisoning AI AND suing millionaire pirates.
And when we have set reasonable boundaries for ML moving forward. People seemed to have missed my statement that I am not opposed to ML only the way it has been implemented and the direction it is moving in.
Add comment